Free app · October 2026 · Runs in your browser
Club Cancellation Risk Scorer
Most cancellations show up in the data weeks before they show up in your inbox. Drop in the member list your club platform already exports and get a ranked list of who is drifting, why, and the fix to offer each one. For wine clubs, mug clubs, bottle clubs, and loyalty programs.
Your file never leaves your computer. No upload, no account, no tracking of what is in it.
Step 1
Drop your club export here.
A CSV from Commerce7, WineDirect, Toast, Square, Shopify, or a spreadsheet. Any columns, any order: the next step matches them. The file stays on your computer. Nothing is uploaded.
How the score works
Every active member starts at zero. Each signal below adds points, the total is capped at 100, and the groups are fixed: 50 and up is High Risk, 25 to 49 is Watch, below that is Healthy. Cancelled members are set aside, not scored. A signal only counts when your export has the column for it, which is why the column step tells you what it could not measure.
- +35Skipped two or more shipmentsNeeds: Skipped shipments
- +30Declined payment, not recoveredNeeds: Declined payments
- +25Paused or on holdNeeds: Status
- +20Skipped one shipmentNeeds: Skipped shipments
- +15No order in nine monthsNeeds: Last order date
- +15Signed up in person, never ordered againNeeds: Sign-up source, plus total orders or orders beyond shipments
- +15Stopped opening emailNeeds: Email opens
- +15Missed the last shipment windowNeeds: Last shipment date (when skips are not in the file)
- +10Never bought beyond shipmentsNeeds: Orders beyond shipments
- +10A year without a visit or eventNeeds: Last visit or event
- +10In the first yearNeeds: Join date
The weights are a starting point drawn from what club managers see most: a skip is the loudest signal, a declined card nobody chased is the quietest and the easiest to fix. Your own cancelled members will tell you the real weights for your club. That is what the deeper version below is for.
What to export
One member list from Commerce7, WineDirect, Toast, Square, Shopify, or whatever runs your club, with as many of these as it offers: name, email, join date, sign-up source (tasting room, online, event), club tier, status, last shipment, skips, declined payments, last order date, total orders, last visit, and recent email opens. Delete any payment card columns before you start. A brewery or distillery runs the same thing on a mug club or bottle club roster. A shop uses its loyalty export. A distributor swaps members for accounts and shipments for reorders.
The deeper version, in Claude
This app scores on fixed rules. The play it comes from goes further: point Claude at your member list and your order history, let it learn from the members who already left, apply those patterns to the active ones, and write a personal save email for each risk group. The full prompt is in the October 12, 2026 issue. The Copy button in the results above gives you a shorter version with your group counts already filled in.
Build your own
Want different weights, your own columns, or a version that lives on your laptop? Paste this into Claude and it builds you a copy. Then tell it what your cancelled members had in common and it tunes the rules to your club.
Build me a single-page web app in one HTML file, no server and no libraries, called Club Cancellation Risk Scorer. It takes a CSV of club members dropped onto the page and lets me match the columns to these fields: member name, email, join date, club tier, status, sign-up source, last shipment date, skipped shipments, declined payments, last order date, total orders, orders beyond shipments, last visit or event, and recent email opens. Any field can be missing. Score every active member (set cancelled members aside) with these points: skipped two or more shipments 35, declined payment 30, paused or on hold 25, skipped one shipment 20, no order in nine months 15, signed up in person and never ordered again 15, zero recent email opens 15, missed the last shipment window 15, never bought beyond shipments 10, a year without a visit 10, in the first year 10. Cap at 100. 50 and up is High Risk, 25 to 49 is Watch, under 25 is Healthy. Show the count in each group, a yearly revenue-at-risk figure (High Risk plus Watch, times shipments per year, times average shipment value, both as inputs), the High Risk members grouped by the main reason they were flagged with a suggested fix that is not a discount, a sortable table of members with their signals, and a Download CSV button. Everything runs in the browser and nothing is sent anywhere. Then change the weights to match my own club: here is what my members who already cancelled had in common in the year before they left: [PASTE WHAT YOU FOUND, OR ASK ME QUESTIONS TO WORK IT OUT].